On Tue, 27 Aug 2019 at 23:40, Philip Rhoades <phil@pricom.com.au> wrote:
Tony,


On 2019-08-28 07:09, Tony Nelson wrote:
> On 19-08-27 08:07:17, Philip Rhoades wrote:
>> People,
>>
>> ...there is a consistent "SERVICE_STOP pid=1 uid=0" command that
>> occurs but I don't know if that is normal or not . .
>   ...
>
> That just means that systemd (pid 1) did something that auditd
> reported.  There are lots of them.  E.g., your:
>
>> Aug  9 22:00:48 localhost audit[1]: SERVICE_STOP pid=1 uid=0
>> auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0
>> msg='unit=systemd-hostnamed comm="systemd"
>> exe="/usr/lib/systemd/systemd" hostname=? a
>
> is about systemd doing something with unit=systemd-hostnamed.

Right - I thought that was probably the case - thanks.

> I have a System Load Monitor applet and a Disk Load Monitor in my
> taskbar, so I can see if memory is filling up or the disk is busy.
 
Right - I have 32GB RAM and the same of swap and I hardly ever see the
swap being used . .

> Still, I think you have some other problem if the system completely
> hangs instead of just getting very very slow.
Yes, me too . . but how to determine what the problem is . . it looks
like whatever it is, messages related to it are not being recorded . .

A rule of thumb is that random failures indicates hardware problems while
reproducible failures indicate software.   These days software problems
often generate log entries, so absence of log entries points to hardware.
Also, many the same hardware failure affects a particular model, so it
is worth searching for problem reports for your model.

I ended up with a "free" ThinkPad that was behaving much as you
describe.  The cooling system was full of dust, but that model was
easy to open up to expose the heatsink, fan, and ducts so I just
cleaned it up and used it for many years.  

Visual inspection sometimes reveals damage that is hard to detect
with software.  I once had a system where networking seemed fine
except NFS was failing -- visual inspection of the network card
revealed fried components.  I've had a number of systems with
failed capacitors (image) or failed connectors (easily disconnected,
some even fell off as system was being moved). 

Use a utility to display temperatures from various
sensors, e.g., lm_sensors

Sniff the power supply for burnt smells.   You may have to take the
cover off the power supply to look for damaged components. 
Inexpensive power supply testers are available.

--
George N. White III